Re: Please test pari-2.17.4 prerelease 1

Dear Bill,

I did my small regression tests against all the WIndows versions listed in your email.

Everything was fine. One thing that I did notice was that there is a behaviour difference between the 32-bit and 64-bit versions. Perhaps a change in memory management?

When I tried

factor(10^67-1)

in the 32-bit version with all the default settings, it worked fine.

In the 64-bit version, I got a stack overflow. I called allocatemem() once and then the factorisation completed successfully.

This behaviour was consistent across all the 32-bit and 64-bit versions in your email.

I am attaching the full session output for both a 32-bit and a 64-bit versions, in case that is helpful.

And of course, you may already be aware of this and this may be expected behaviour.

Regards,

Paul

On Mon, 22 Jun 2026 at 13:07, Bill Allombert <Bill.Allombert@math.u-bordeaux.fr> wrote:

Dear PARI developers,

We have made available a prerelease of PARI 2.17.4 (stable).

Please test to make sure there are no regressions!

The expected release date is 29/06/2026, but it might be delayed because
I am missing MacOS binaries for Arm-based Mac (M1, M2, M3, etc.)

Please test the prerelease tarball:
<https://pari.math.u-bordeaux.fr/pub/pari/snapshots/pari-2.17.4-pre1.tar.gz>

The digital signature is
<https://pari.math.u-bordeaux.fr/pub/pari/snapshots/pari-2.17.4-pre1.tar.gz.asc>

Please test the 64bit Windows installer:
<https://pari.math.u-bordeaux.fr/pub/pari/windows/snapshots/Pari64-2-17-4-pre1.exe>

I have also built the 32bit Windows installer:
<https://pari.math.u-bordeaux.fr/pub/pari/windows/snapshots/Pari32-2-17-4-pre1.exe>

and the following standalone Windows binaries:

32bit
<https://pari.math.u-bordeaux.fr/pub/pari/windows/snapshots/gp32-2-17-4-pre1.exe>
<https://pari.math.u-bordeaux.fr/pub/pari/windows/snapshots/gp32-readline-2-17-4-pre1.exe>

64bit
<https://pari.math.u-bordeaux.fr/pub/pari/windows/snapshots/gp64-2-17-4-pre1.exe>
<https://pari.math.u-bordeaux.fr/pub/pari/windows/snapshots/gp64-readline-2-17-4-pre1.exe>

64bit, multithreaded
<https://pari.math.u-bordeaux.fr/pub/pari/windows/snapshots/gppthread64-2-17-4-pre1.exe>
<https://pari.math.u-bordeaux.fr/pub/pari/windows/snapshots/gppthread64-readline-2-17-4-pre1.exe>
I will provide link to MacOS binaries later.

For intel-based mac:

the following MacOS DMG:

<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/PariGP-2.17.4-pre1.dmg>
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/PariGP-2.17.4-pre1-pthread.dmg>
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/PariGP-full-2.17.4-pre1.dmg>
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/PariGP-full-2.17.4-pre1-pthread.dmg>

the following MacOS stand-alone binaries:
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/gp-2.17.4-pre1-osx>
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/gp-2.17.4-pre1-pthread-osx>

For arm-based ("Apple silicon") mac
(not available yet)

<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/PariGP-arm-2.17.4-pre1.dmg>
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/PariGP-arm-2.17.4-pre1-pthread.dmg>
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/PariGP-full-arm-2.17.4-pre1.dmg>
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/PariGP-full-arm-2.17.4-pre1-pthread.dmg>

the following MacOS stand-alone binaries:
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/gp-arm-2.17.4-pre1-osx>
<https://pari.math.u-bordeaux.fr/pub/pari/mac/snapshots/gp-arm-2.17.4-pre1-pthread-osx>

On behalf of the PARI Group,
Bill
P.S. the CHANGES

Bug numbers refer to the BTS at http://pari.math.u-bordeaux.fr/Bugs/

[last column crossreferences current development release 2.18.1]
Done for version 2.17.4 (released 29/06/2026):
Fixed
BA 1- call(printsep, [" "]) -> SEGV [F44]
2- forsubset: handle trivial case of negative cardinalities [#2652] [F47]
BA 3- ellchangecurve(E).roots could be wrong [#2660] [F48]
4- charpoly(Mod(x+y,y^2+x+1),y) -> invalid t_POL [#2668] [F50]
5- writebin(file) only saved polynomials variables in file [F53]
BA 6- lfun(1/3+I*10^19) -> overflow in expo() [F54]
BA 7- fflog(a, b, o) was very slow for small o but large fields [#2676] [F55]
LGr 8- psi(1+O(x^5)) -> wrong result [#2677] [F56]
9- zetahurwitz(large s, |x| > 1) is inaccurate (correct absolute accuracy
but zeta(s,x) is small and relative error is large) [#2673] [F57]
BA 10- ellcard(ellinit([ffgen([53,36]),1])) -> wrong result [F58]
11- polcyclofactors(((x^6+x^3+1)*(3*x^3+1))^2) -> oo loop [#2680] [F59]